another 34 per cent of the hens (0.43 times 0.79) were also successful. 
This gives a total of 91 per cent. If only half of the 43 per cent 
renested, the total percentage of successful females would be 7h per 
cent. Using Kalmbach's statistics in this fashion (and they were not, 
of course, calculated solely for mallards), we can now see the range 
of age ratios possible at this stage of the life cycle: 
' Hypothesis Data 
(1) If no females renest, there 66 pr. x 7.0 = 62 young 
will be about 2.3 juveniles 462:200 adults::2.3:1 
per adult 
(2) If half renest, the ratio is 57 pr. x 7.0 = 399 young 
2.6 to 1 23-5 x 0.79 x 7.0 = 130 young 
5293200: 32.631 
(3) If all unsuccessful the first 57 x 7.0 = 399 young 
time renest, then the age 43 x 0.79 x 7.0 = 238 young 
ratio is 3.2 to 1 63732003 : 3.231 
These estimates are minimal values because they do not consider mor- 
tality in the adults during the nesting period. If this were 25 per 
cent and we corrected (2) above by using 150 adults in lieu of 200 
in the proportion, the age ratio in such a population would then be 
of the order of 3.5 juveniles to each adult. If itwere 10 per cent, 
the age ratio would be 2.9. This review of some possible age ratios 
in mallards will, I hope, emphasize that high brood counts do not 
necessarily imply high ratios of young to adults in mallard popula- 
tions. An enormous fund of information on age ratios in this species 
has been accumulated during the hunting season, but a critical analysis 
of these data has not yet been published. 
The sex ratio for adult and subadult mallards appears to be 
close to an even one. McIlhenny (190) has reported trapping 1728 
males and 1713 females from 193) to 1938 in Louisiana where the 
birds winter. Ratios approaching this one have also been directly 
observed by Beer (195) in southwestern Washington and by Erickson 
(1943) in Minnesota. Mallards having an unbalanced sex ratio (61.3 
per cent males to 38.7 per cent females in a sample of 9783) have been 
trapped in northeastern Illinois from 190 to 19,5 (Mann, Thompson, 
and Jedlicka 1947); but trapping data in that state seem to be biased 
(Hawkins, personal communication). The preponderance of males some- 
times noted in hunters' bags may not be significant (Hawkins, Bell- 
rose, and Smith 196). 
Banding work in North America 
Since 192), about 275,000 mallards have been banded in North 
America, mostly in subadult or adult stages, and usually during the 
hunting season (September to January or February). Critically com- 
petent age determinations have been used by relatively few banders 
after September 1 in any given year. The most notable banding opera- 
tions carried out in this connection have been the work of the Game 
65 
